Volume 18, No. 6, 2021
An Ensemble Machine Learning Techniques With Dolphin Swarm Algorithm For COVID-19 Sentiment Analysis
Vathsala.M.K. , Sharmila Suttur C
Abstract
The range of Covid-19 has raised concerns for the health of people all across the world. When faced with a catastrophe like a natural disaster, political turmoil, or terrorism, social scientists and psychologists want to know how individuals express their feelings and sentiments. There have been numerous psychological concerns caused by the COVID-19 epidemic, including depression in light of recent social changes and a lack of jobs. News and thoughts regarding it are rapidly being shared via social media. To make the best use of available resources, a realistic appraisal of the current situation is required. We use an ensemble ML technique to do sentiment analysis on Covid19 tweets in this study. To effectively deal with the current pandemic crisis, identification of Covid-19 attitudes from tweets is necessary. Text Blob is used to extract sentiments from the dataset after it has been cleaned up using pre processing procedures. Using our suggested feature set, the Dolphin Swarm Algorithm(DSA), we evaluated the performance of various machine learning classifiers. Positive, neutral, or negative tweets are all considered to be equal. Classifiers are judged based on their accuracy, precision, recall, and F1 score, among other metrics. A 90.51 percent accuracy score using our proposed concatenated feature set shows that Support Vector Machine (SVM) Classifiers outperform all other models. Compared to ML classifiers, the Multi-layer perceptron (MLP) has poor accuracy.
Pages: 1816-1833
Keywords: COVID-19; Dolphin Swarm Algorithm; Machine Learning; Multi-Layer Perceptron; Sentiment Analysis